Electronic systems, such as, for example, computer systems include several integrated circuit (IC) devices that generate heat during operation. For effective operation of the computer system, the temperature of the IC devices have to be maintained within acceptable limits. While the problem of heat removal from IC devices is an old one, this problem has increased in recent years due to greater numbers of transistors that are packed into a single IC device while reducing the physical size of the device. Increasing number of transistors compacted into a smaller area results in a greater concentration of heat that must be removed from that smaller area. Bundling multiple computer systems together, such as, for example, in a server, further aggravates the heat removal problem by increasing the amount of heat that has to be removed from a relatively small area.
In a typical computer server (“server”), multiple computer server modules (“modules”) are stacked together in a rack or a case to consolidate network resources and minimize floor space. Modules which are designed for use in a server configuration are typically characterized by a motherboard comprising heat generating electronic components (such as IC devices) housed in a modular chassis or case, which in turn is mounted together with other similar modules, in a rack, blade cabinet, blade server, or other support structure. In practice, multiple servers (each comprising several modules) are typically located in an enclosed space such as a server room or a data center. During operation, the electronic components in the individual modules generate heat which must be removed for effective functioning of the server. FIG. 1 illustrates an prior art method used to cool multiple servers (each containing multiple modules) housed in an enclosed environment such as, for example, a server room. In such a prior art system, cooling fans are used circulate ambient air from the server room through the multiple modules of a server to absorb heat therefrom. In the prior art system, cool air directed into the server room through a cold air plenum is passed through the servers to absorb heat generated by IC devices and other heat generating components therein. After absorbing the generated heat, the heated air is exhausted back into the server room. This heated air is directed through a warm air plenum to a computer room air conditioning (CRAC) system to cool the air and recirculate it back to the server room through the cold air plenum.
It is known that a large portion (greater than about 31%) of the energy consumption of a typical server room is used in the operation of the CRAC system, and that significant energy savings and resultant green house gas reduction can be achieved by improving the efficiency of the CRAC system. “Data Center Energy Characterization Study Site Report,” February 2001, available at http://hightech.lbl.gov/documents/DATA_CENTERS/DC_Benchmarking/Data_Center_Facility1.pdf; “Energy Consumption of Information Technology Data Centers,” and references cited therein, Iyengar et al., December 2010, available at http://www.electronics-cooling.com/2010/12/energy-consumption-of-information-technology-data-centers/. Improving the cooling efficiency of servers housed in a server room thereby enables more efficient utilization and conservation of available energy resources, and green house gas emission reduction.
The disclosed liquid cooling systems and methods are directed to an energy efficient approach of cooling one or more servers located in an enclosed environment, such as a server room.